Short Courses

The VGC 2023 conference will provide a valuable opportunity to participate in short courses aligned with the conference theme. Hosted in Freiberg and led by top experts in the field, these courses offer the latest techniques and insights for attendees to enhance their knowledge and skills.

The course offerings have been thoughtfully categorized into three distinct areas, with each category comprising of two 3-hour sessions:

      • Workshop Stream 1: Hyperspectral imaging, correction and analysis
      • Workshop Stream 2: Processing point clouds with Python
      • Workshop Stream 3: Analysing point clouds in 3- and 4-D

The sessions will run in parallel:

 

Hyperspectral imaging, correction and analysis

Processing point clouds
with Python

Analysing point clouds in 3- and 4-D
9:00 – 10:00 Tea / Coffe
10:00 – 13:00 Hyperspectral image acquisition

Automated processing and interactive visualization of dense point-cloud data using python

 Digital outcrop modelling with CloudCompare, Gempy and LiquidEarth
13:00 – 14:00 Lunch (included in the short courses fee)
14:00 – 17:00 Hyperspectral correction and analysis using hylite and hyperclouds Semantic segmentation and surface reconstruction from point cloud data

4D change analysis of near-continuous LiDAR time series for applications in geomorphic monitoring 
Click on the titles to see a detailed description of the courses:

Workshop Stream 1: Hyperspectral imaging, correction and analysis

Hyperspectral image acquisition

by Moritz Kirsch, Sandra Lorenz

Hyperspectral imaging is a broadly applied technique for remotely characterizing materials and estimating e.g., mineralogy. This session will cover the basics of hyperspectral remote sensing theory, and then introduce a variety of ground and UAV-based hyperspectral sensors. Data acquisition will then be demonstrated using several tripod mounted VNIR-SWIR and LWIR sensors.

Prerequisites: Interest in remote sensing or hyperspectral imaging 

Hyperspectral correction and analysis using hylite and hyperclouds

by Sam Thiele, Sandra Lorenz, Moritz Kirsch

After acquiring a hyperspectral image, a variety of correction, fusion and analysis steps are needed to derive accurate data. In this session, we will use the open-source python package hylite to (1) accurately locate a hyperspectral sensor relative to a digital outcrop, (2) use the geometric information from the digital outcrop to apply atmospheric and topographic corrections, (3) back-project corrected spectra to derive a hypercloud, and (4) extract mineralogy information.

Prerequisites: Basic python programming, jupyter notebooks

Workshop Stream 2: Processing point clouds with Python

Automated processing and interactive visualization of dense point-cloud data using python

by Florent Poux

This session will introduce open-source python tools for processing of point cloud data, including the development of automated pipelines for point cloud subsampling, structuring, denoising, filtering etc. Open source tools for the visualization and sharing of large point cloud datasets will also be demonstrated, and approaches for interacting

Prerequisites: Basic python

Semantic segmentation and surface reconstruction from point cloud data

by Florent Poux

Semantic segmentation is key for many types of point cloud analyses (e.g., segmentation of different objects, land cover classification, geological mapping). This session introduces and applies several different machine learning approaches for semantic segmentation using open-source python machine learning tools, and explores associated data representations (meshes, voxels, etc.) and their respective advantages and limitations.

Prerequisites: Basic python

Workshop Stream 3: Analysing point clouds in 3- and 4-D

Digital outcrop modelling with CloudCompare, Gempy and LiquidEarth

by Sam Thiele, Miguel de la Varga and Simon Virgo

While the creation of digital outcrop models is now relatively straight-forward, the identification, mapping and extrapolation of geological structures captured in such models  remains challenging. In this session we will demonstrate a workflow that uses the Compass plugin in CloudCompare to extract structural information from a digital outcrop model, and then derive a 3D geological model and associated classified point cloud (geological map) using the open-source python package gempy. Novel AR and VR tools for interacting with combined digital outcrop and 3D geological models will also be demonstrated.

Prerequisites: Interest in digital outcrop geology, basic python (useful but not essential) 

4D change analysis of near-continuous LiDAR time series for applications in geomorphic monitoring

by Katharina Anders and Roderik Lindenbergh

This short course will introduce the acquisition and processing of 4D point clouds (theory), and present methods to handle, explore, and visualize the data (hands-on) in a first part. The second part will feature time series-based methods of analyzing 4D point clouds (theory and hands-on), demonstrating how the temporal information in these data can be used to identify and assess occurrences and patterns of different surface activities in a scene. Notably, we will introduce time series clustering and the extraction of 4D objects-by-change. The test dataset will be a laser scanning time series of a sandy beach. Hands-on will be performed with Python (basic skills required), where methods are available in the open-source library py4dgeo.

Prerequisites: basic python (useful but not essential)